Events associated with the Kennedy assassination offer a compelling case study regarding obsolete data formats and digital preservation.
Shortly after the assassination of President Kennedy on this day 48 years ago, an organization turned to the latest computer technology in an effort to study the tragedy. From November 26 through December 3, 1963, the National Opinion Research Center at the University of Chicago conducted a survey of citizen reactions to the event. The results were recorded on paper punch cards, which were used to input data into the mainframe computer used to tabulate study data. Summary results were then published and time moved on.
When another national catastrophe stuck on September 11, 2001, NORC researchers acted to replicate the 1963 study by asking the same kind of questions to assess public reaction. The aim was to compare how the nation responded to two very different tragedies. This in turn raised the question: where was the original Kennedy assassination data set? NORC had some summary printouts and tapes, but the complete data–the punch cards–were unaccounted for. Without the cards, researchers could not make the most meaningful comparisons. NORC archivists set out on a hunt for the elusive data, and after reviewing old indices to thousands of boxes in a private storage facility, the original punch card decks were found.
Celebration had to wait. NORC had stopped using punch cards a quarter of a century earlier and had to arrange for a third party to convert the Kennedy cards to a format that current computers could read. A facility in New York could do the work, but not without causing some anxiety. There was only one copy of the cards, which meant there was no room for error, either in getting the cards to the facility or in processing them. Nerves spiked when the facility reported they had to refurb our punched card equipment, it had been sitting around so long it got a little rusty.
In the end, all worked well and the data set was successfully migrated to a modern data format.
While this particular foray into data archeology had a happy ending, it illustrates the hurdles that time and obsolescence can present to keeping digital data alive and useful.
Ideally, the original Kennedy assassination data set should have been continually moved forward to new generations of media and file formats. Not doing so left the data set at the mercy of a distant and rusty mechanical relic.
I’m happy to say that this perspective was amplified and highlighted through the NDIIPP-supported Data-PASS project, a voluntary partnership of organizations created to archive, catalog and preserve data used for social science research.